Per-Entity Metadata Working Group

["Cantor, Scott" <>]

On 8/10/16 7:01 PM, Patrick Radtke wrote:
>
> From my perspective uptime measurements on their own shouldn't drive
> the decision since they don't take into account user impact. A daily
> outage of 20 seconds at 4 am may go unnoticed, while 2 hour outage one
> day a year will result in a lot of complaints even though they both
> have the approximately the same uptime measurements for the year.

You can't assume that distribution. If you're telling me you're willing
to be down a minute a week or more, I have to assume that minute will be
at the worst possible time, and I can't have that. So anything short of
5 9s doesn't mean much to me. I think that sends me a message that I'd
better implement more caching.

> For 1) and 2) the failures at a specific CDN location would likely
> affect a small portion of the users.

But that's like saying the downtime metric doesn't matter. You can't
know which users, how important they are, etc. That could be a fatal,
game-changing outage.

> I know we don't want to put intelligence into the client since the
> client base is too heterogenous, but that is the way Netflix went with
> all their internal services to achieve a highly perceived uptime.

That's what 4 9s says to me, and that's what I said on the call.

Ultimately most of those "clients" today can't do metadata at all and
probably never will, and the rest will be static. So the actual set of
clients impacted here is not large. I just need to know the requirements.

-- Scott